CI/CD is now AI/CD
summary
AI multiplies code output so CD becomes the bottleneck; the concrete proposal routes each change by measured risk using CD policy ops that check provenance, signatures (cosign, SLSA), tests and SBOMs, then return a decision to a separate executor (Octopus, Argo CD/Flux, Argo Rollouts) and telemetry (OpenTelemetry, Prometheus, Grafana) for rollback. Examples: a small UI tweak takes a fast automated canary path, payment-logic changes take a controlled human-verified staged path. The architecture proves feasible in examples and is designed to earn — and revoke — automation based on evidence.
AI-written summary · about
chapters (13)
- 0:00Introduction to AICD
- 1:14CI/CD fundamentals and bottlenecks
- 3:24CD policy ops concept
- 6:35Determining deployment confidence and risk
- 8:41Policy engine and execution separation
- 10:26Architecture and tooling examples
- 11:49AI as an operator in the loop
- 13:14Evidence layer and security
- 13:59Low-risk change automation
- 15:06High-impact change handling
- 17:12Recovery and feedback loops
- 18:15Earning autonomy over time
- 19:38Q&A on AI/CD impact tracking
related talks
08mCI/CD Is Now AI/CD
23mCI/CD: Conscious Uncoupling
19mRethinking CI & CD for Agentic Workflows
22mCI and CD as a Service: A multi-vendor driven system
33mCI/CD for Autonomous Drive and other complex systems - What we have learned about...
29mCI/CD as a Product
42mCI/CD, is the Divorce Inevitable?
17mCI/CD - Build/Deploy Automation